Meta Interview Questions

Meta System Design Interview Questions

Practice 1,159 real Meta interview questions for 2026. Covers top categories — Coding & Algorithms, Analytics & Experimentation, Data Manipulation (SQL/Python), Behavioral & Leadership, and System Design — across Software Engineer, Data Scientist, Machine Learning Engineer, Data Engineer, and Product Manager roles. Real questions from actual interviews with detailed solutions. Expect a software-engineering-heavy loop: timed algorithmic coding (trees, arrays, graph/maze problems, delimiter/CSV parsing), system-design prompts like leaderboards, flight search and online-judge architectures, and an increasingly common AI-assisted coding round that mirrors real workflows. Data Scientist rounds emphasize product analytics and experimentation—designing tests, diagnosing spend drops and bots, evaluating unconnected content, and writing SQL for multi-account, seller, and vehicle metrics. Machine Learning Engineer questions skew toward recommender and ranking work (place and friend recommendation, sparse-matrix ops, linear-regression derivations, newsfeed dislike models). Data Engineers focus on data modeling, ETL, capacity calculations, reservations/utilization queries, and production SQL/Python tasks. For interview preparation, prioritize timed coding practice, system-design templates, rigorous SQL drills (joins/CTEs/aggregation), clear A/B-testing frameworks, and concise STAR behavioral stories tied to measurable impact.

1.2k Questions 1 Company07.06.2026
Showing 20 results
Role
Meta logo
Meta
Medium
Data Scientist

Join datasets and compute conversion by assignment

You are given two CSVs. Create tables and write SQL to produce both visit-level and visitor-level conversion datasets, then aggregate conversion by as...

Data Manipulation (SQL/Python)
0
0
5 people solved
Oct 13, 2025
Meta logo
Meta
Medium
Software Engineer

Solve BST range average and merge intervals

1) Given a binary search tree and an inclusive range [L, R], compute both the sum and the average of all node values where L <= val <= R. Use the BST ...

Coding & Algorithms
5
0
52 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Data Engineer

Compute capacities after site closures

You are given a nested dictionary redistribution where redistribution[closed_site][dest_site] equals the additional capacity required at dest_site if ...

Coding & Algorithms
6
0
70 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Software Engineer

Implement string dedup and mirror tree traversal

- String cleanup with group deletions (LeetCode-inspired): Given a string s consisting of lowercase letters, repeatedly delete any maximal contiguous ...

Coding & Algorithms
1
0
26 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Software Engineer

Implement BST vertical traversal and list conversion

Given the root of a binary search tree (BST): 1) Implement vertical order traversal: assign the root column 0, left child column −1, right child colum...

Coding & Algorithms
2
0
46 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Software Engineer

Compute left and right views once

Given the root of a binary search tree, output two lists: the nodes visible from the left side and the nodes visible from the right side, from top to ...

Coding & Algorithms
4
0
32 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Machine Learning Engineer

Find kth smallest in sorted matrix

Given an n x n matrix where each row and each column is sorted in nondecreasing order, and an integer k (1 ≤ k ≤ n^ 2), return the k-th smallest eleme...

Coding & Algorithms
2
0
42 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Software Engineer

Compute longest rising path in a grid

Given an m×n grid of integers, find the length of the longest path where each step moves up, down, left, or right to a strictly larger value. Return t...

Coding & Algorithms
3
0
42 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Software Engineer

Fill rooms with nearest gate distances

You are given an m×n integer grid representing rooms, where -1 is a wall, 0 is a gate, and INF (a large sentinel such as 2^31− 1) is an empty room. Fo...

Coding & Algorithms
1
0
26 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Machine Learning Engineer

Find connected region sizes in matrix

Given an m x n grid of characters grid and a target character ch, implement: ( 1) sizeOfRegion(grid, r0, c0, ch) -> int that returns the size of the 4...

Coding & Algorithms
5
0
57 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Software Engineer

Solve grid path and robot navigation

Part A — Shortest path in a grid with obstacles: Given an m×n grid of 0/1 cells (0 = open, 1 = obstacle), start at (0, 0) and end at (m−1,n− 1). You m...

Coding & Algorithms
4
0
33 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Software Engineer

Implement power and balance-parentheses algorithms

Design and implement two algorithms: 1) Fast exponentiation: Implement a function pow(x, n) that returns a real number x raised to an integer exponent...

Coding & Algorithms
2
0
31 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Software Engineer

Count subarrays equal to target

Given an integer array nums and an integer k, return the number of contiguous subarrays whose sum equals k. Provide an algorithm better than O(n^ 2), ...

Coding & Algorithms
2
0
23 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Data Engineer

Optimize SQL to minimize scans

Given a large analytics query, refactor it to minimize table scans. 1) Replace unnecessary CTEs that cause multiple scans with inline aggregations or ...

Data Manipulation (SQL/Python)
3
0
6 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Software Engineer

Optimize flight costs and find grid path

1) Given two integer arrays D and R of equal length n, where D[i] is the cost to depart on day i and R[j] is the cost to return on day j, choose indic...

Coding & Algorithms
4
0
30 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Machine Learning Engineer

Find kth smallest pair sum with heaps

You are given two nondecreasing arrays A (length n) and B (length m) of non-negative integers and an integer k (1 <= k <= n*m). Return the k-th smalle...

Coding & Algorithms
3
0
48 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Software Engineer

Find top-K frequent values with tiebreaks

Question Given an integer array nums and an integer k, return the k values with the highest frequency. The array can contain up to 10^6 elements. When...

Coding & Algorithms
5
0
40 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Software Engineer

Merge two interval lists into a union

You are given two lists of closed intervals [start, end]. Each list is individually sorted by start and contains non-overlapping intervals. Merge the ...

Coding & Algorithms
1
0
23 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Software Engineer

Design LCA and find K closest points

Part A — LCA with parent pointers: You are given two nodes a and b in a rooted tree where each node has a parent pointer (you may or may not have dire...

Coding & Algorithms
2
0
24 people solved
Sep 6, 2025
Meta logo
Meta
Medium
Machine Learning Engineer

Compute binary tree left-side view

Given a binary tree, return the values visible from the left side when looking from top to bottom. If multiple nodes exist at the same depth, the left...

Coding & Algorithms
4
0
50 people solved
Sep 6, 2025

Frequently Asked Questions

How difficult are Meta interview questions?
Meta interview questions span a wide difficulty range because they must screen candidates from entry to senior levels across many functions. Expect coding rounds to map to medium-to-hard algorithmic problems that appear in top 100 problem lists for software engineers, and expect data roles to face challenging SQL, experiment diagnosis, and product-analytics problems that require clean metric definitions. Machine learning roles emphasize recommendation and ranking tradeoffs and model complexity, while data engineers encounter large-scale ETL and modeling puzzles. Difficulty scales with level: entry hires see clearer, bounded problems; senior hires face ambiguous tradeoffs and system-wide thinking.
What is Meta's interview process and where do these questions appear?
Meta typically runs a multi-stage process: recruiter screen, one or two technical screens or an online assessment, a full loop of onsite-style interviews, then debrief, committee review, and offer. The full loop mixes coding, system or product design, role-specific technical rounds, and behavioral interviews. Software-engineer candidates spend most time on coding and design; data scientists focus on SQL, experimentation, and product analytics; machine-learning engineers see modeling and recommendation design; data engineers handle SQL, data modeling, and pipeline questions; PMs get product-design and analytics probes. In 2025–2026 some teams pilot AI-enabled coding rounds.
How should I structure a preparation timeline for a Meta interview?
A focused six-week plan works well: weeks one and two cover fundamentals—data structures, algorithms, SQL basics, and experiment design; weeks three and four emphasize timed problem practice, mock phone screens, and role-specific cases (A/B diagnosis for data scientists, model design for MLEs, ETL modeling for data engineers); week five concentrates on system or product design and behavioral storytelling; week six is for full mock loops, timing, and refining communication. Practice with realistic tools, simulate loop pacing, and schedule a debrief after each mock to iterate on clarity, edge-case handling, and time management.
Which technical subtopics are most commonly tested for each role at Meta?
For Data Scientist interviews the recurring technical themes are product-metric definition, diagnosing experiment and spend drops, counting multi-account interactions, SQL for multi-entity metrics, and ranking or recommendation evaluation such as shop ad ranking. Software-engineer questions frequently focus on timestamped state and versioned systems, leaderboards and ranking, maze/graph traversal and tree/array transforms, delimiter and CSV parsing, and scalable search or flight-search style designs. Machine-learning engineers see place and friend recommendation design, sparse-matrix operations, ranking/loss choices, and feed dislike or personalization models. Data engineers repeatedly face entity modeling for feed and booking data, SQL analytics for utilization and reservations, and capacity-aware aggregation challenges.
What standout tips and common pitfalls should I watch for in Meta interviews?
Start interviews by clarifying requirements and expected outputs, then propose measurable success metrics; this prevents misaligned solutions. For coding, think aloud, handle edge cases, state complexity up front, and write a couple of quick tests. In design rounds quantify load, storage, and tradeoffs rather than vague features. Data roles must define metrics, guardrails, and experiment assumptions before jumping to analysis; common pitfalls are ambiguous metric definitions, peeking at tests, and ignoring instrumentation limits. For AI-assisted coding rounds, use the assistant to accelerate boilerplate but validate logic and corner cases yourself. Finish each answer with a concise summary of impact and tradeoffs.

Explore more Meta interview questions

Jump straight to Meta questions for a specific role or category.

By role
By category
In-depth guides
Across all companies

Featured Meta interview prep guides

Concept walkthroughs, worked examples, and the real questions from candidate reports.

Editorial prep
Software Engineer
Meta interview
Read the guide
Editorial prep
Data Scientist
Meta interview
Read the guide